A quarterly license review catches cost drift while it is still small, and buyers who run one cut renewal surprises by 30 to 50 percent compared with those who only look at an estate when a contract is about to expire. The math is simple. Software entitlements decay against reality every day that people are hired, projects end, environments spin up, and modules go unused. A review you run four times a year finds that drift in increments you can fix. A review you run once, in the panic before a renewal, finds it all at once, on the vendor's timeline, with no room to act.
This page sets out a working quarterly cadence: the four reviews, who owns each one, the data each needs, and the decisions each is supposed to produce. It is the operating rhythm that sits under every other control in a license compliance program, and it is the single habit that most separates estates that negotiate from a position of knowledge from estates that negotiate from surprise.
Inside This Guide
Why quarterly, not annual
An annual review is roughly 90 days too late by design. Most enterprise agreements give a vendor the right to true up under-licensed usage, and most renewals carry an uplift the buyer has to actively contest. Both are time-sensitive. A finding that surfaces 11 months into a 12-month cycle leaves no runway to right-size, to reallocate, or to build the alternative that gives a negotiation its strength. The same finding surfaced in a quarterly pass leaves three quarters to act.
Quarterly also matches how cost actually accumulates. Seat counts climb with hiring. Cloud consumption climbs with every new workload. Module adoption spreads as teams discover features that sit outside the entitlement. None of these move once a year; they move continuously, and a quarterly check turns a continuous problem into four discrete, manageable decisions. The cadence does not need to be heavy. It needs to be regular, owned, and tied to a small set of numbers that mean something.
Q1: the entitlement baseline
The first review of the year rebuilds the effective license position for every material vendor, so the rest of the year measures against a clean baseline. This is the reconciliation of what you are entitled to against what you have deployed, vendor by vendor, metric by metric. It is the slowest of the four reviews and the most important, because every later decision assumes the baseline is right.
The output is a position statement per vendor: licensed quantity, deployed quantity, the gap in each direction, and a confidence rating on the data. Over-deployment is compliance exposure to remediate. Under-deployment is shelfware to harvest or shed. A low confidence rating is itself a finding, because a vendor's measurement will fill any gap your own measurement leaves open.
Q2: the consumption and shelfware review
The second review answers one question: what are you paying for that nobody uses? Shelfware runs 10 to 30 percent of seats in a typical estate that has grown by acquisition or reorganization, and it is the easiest spend to recover because removing it changes nothing operationally. Pull last-login and feature-usage data for every per-seat product, flag any account dormant for 90 days, and reconcile the active count against the licensed count. The detail of how to act on this sits in our right-sizing licenses guide.
For consumption-based products the same review tracks the run rate against the commitment. A committed-spend deal that is running hot needs a forecast conversation before it breaches; one running cold needs a reallocation plan before the term closes and the unused commitment is simply lost. Both are findings you want in Q2, not at renewal.
The 90-day dormancy rule. Any per-seat account with no login for 90 consecutive days should auto-flag for reclaim review. In estates we have measured, that single rule surfaces enough recoverable seats to fund the entire review program several times over, and it removes the awkward conversation of justifying a seat nobody can name a user for.
Q3: the renewal-runway review
The third review looks forward 12 months at every renewal on the calendar and asks whether each one has enough runway to be negotiated rather than rubber-stamped. A renewal needs 6 to 9 months of preparation to be negotiated well: time to build the position, to benchmark the price, to model the alternatives, and to let the vendor's fiscal pressure work in your favor. A renewal that surfaces with 60 days left is a renewal you will pay close to list for.
For each upcoming renewal, the review sets the target outcome, identifies the benchmark, names the alternative that anchors the talks, and assigns an owner. Where the contract carries an uplift cap or a price-hold clause, the review confirms it still applies. Where it carries an auto-renewal trigger, the review confirms the cancellation notice window has not already opened. The vendor-by-vendor tactics that turn this runway into price belong to our negotiation playbook by vendor.
Q4: the budget and benchmark review
The fourth review reconciles the year against the budget and benchmarks the estate against the market before the next planning cycle locks numbers in. This is where you measure discount erosion: a discount band that read 45 percent two years ago and reads 38 percent now has eroded 7 points, and that erosion compounds every renewal it goes unchallenged, a pattern we cover in discount erosion at renewal. Benchmarking the realized price against comparable deals, using a defensible method, gives the next year's negotiations their anchor; our contract benchmarking methodology sets out how to do that without relying on a single vendor's selective data.
Q4 also closes the loop on the year's findings. Every flag raised in Q1 through Q3 either resulted in an action or it did not, and the review records which, so the program improves rather than repeating the same misses. The output feeds the budget, so next year's number reflects the right-sized estate, not last year's drift carried forward.
The data each review needs
The cadence is only as good as the data behind it, and four data sources cover the great majority of what the reviews need. The table below maps each review to its inputs and its primary output.
| Review | Primary inputs | Decision produced | Typical owner |
|---|---|---|---|
| Q1 baseline | Entitlement records, deployment scans, contract terms | Effective license position per vendor | SAM / IT asset management |
| Q2 consumption | Login and feature usage, consumption run rate | Shelfware reclaim and reallocation list | SAM with application owners |
| Q3 renewal runway | Renewal calendar, contract clauses, market benchmarks | Negotiation plan per renewal | Procurement / sourcing |
| Q4 budget and benchmark | Realized prices, discount history, peer benchmarks | Budget input and benchmark anchors | Procurement with finance |
Notice that ownership rotates between asset management and procurement across the year. That is deliberate. The data reviews belong to the team that holds the measurement; the commercial reviews belong to the team that holds the contract. A cadence that puts all four on one team either starves for data or starves for commercial judgment.
Metrics that signal trouble
A review needs a small set of numbers that trip an alarm, so attention goes where it is needed rather than spreading evenly across an estate where most of it is fine. Five metrics carry most of the early-warning value: the share of per-seat accounts dormant past 90 days, the gap between consumption run rate and commitment, the months of runway on the nearest material renewal, the realized discount versus the prior renewal, and the count of deployments that fall outside a known entitlement.
Each of these has a threshold that should trigger an action rather than a note. Dormant seats above 10 percent trigger a reclaim. Consumption tracking 15 percent off forecast triggers a model rebuild. A renewal inside 6 months without a plan triggers an escalation. A discount that has slipped more than 5 points triggers a benchmark. A deployment outside entitlement triggers a position recheck before a vendor turns it into a claim, the same exposure we describe in indirect access risk.
Who owns the cadence
A cadence with no owner becomes a calendar invite nobody runs, so the program needs a single accountable owner and a clear split of responsibilities. The accountable owner is usually the software asset management lead or a sourcing manager who reports the findings up; the executive sponsor is the IT finance leader who acts on them. The two data reviews belong to asset management because that team holds the measurement tooling and the entitlement records. The two commercial reviews belong to procurement because that team holds the contracts and the vendor relationships. Application owners feed both, because only they can confirm whether a flagged seat or tier is genuinely surplus.
The reporting line matters as much as the ownership. A review that produces findings nobody is required to act on degrades into a quarterly note, so each review should close with named actions, owners, and due dates that carry into the next quarter's agenda. Tie the program to the budget cycle so the Q4 benchmark feeds the next year's number directly, and give the executive sponsor a one-page summary each quarter showing recovered spend, open exposures, and renewals on the horizon. That single page is what keeps the cadence funded, because it turns an operational habit into a visible financial result that leadership can see and defend.
Right-size what the cadence covers to the estate's materiality. The largest vendor relationships deserve all four reviews in full; smaller relationships can run a lighter version, a single combined review twice a year, so the program's effort tracks the money at stake. The goal is coverage proportional to spend, not uniform effort across an estate where most of the risk sits in a handful of contracts. A cadence that spends equally on a $5M Oracle relationship and a $50K utility tool is misallocating the one resource the program actually has, which is attention.
Run these four reviews on a standing calendar, assign each an owner and a threshold, and the estate stops surprising you. The drift that used to arrive once a year as an invoice arrives four times a year as a decision you still have time to make. That shift, from surprise to decision, is the entire point of the cadence, and it is why a quarterly rhythm is the cheapest and highest-return control most software estates do not yet have. When the reviews surface findings that need a market view or a counterpart at the vendor, our software licensing advisory and vendor audit defense teams run the reconciliation and the negotiation with you.